FRIB Project Description

0329-EX-PL-2016 Text Documents

Michigan State University

2016-05-11ELS_176661

MOB01 Proceedings of HIAT 2012, Chicago, IL USA THE FRIB PROJECT – ACCELERATOR CHALLENGES AND PROGRESS* J. Wei#1, D. Arenius2, E. Bernard1, N. Bultman1, F. Casagrande1, S. Chouhan1, C. Compton1, K. Davidson1, A. Facco1,4, V. Ganni2, P. Gibson1, T. Glasmacher1, L. Harle1, K. Holland1, M. Johnson1, S. Jones1, D. Leitner1, M. Leitner1, G. Machicoane1, F. Marti1, D. Morris1, J. Nolen1,3, J. Ozelis1, S. Peng1, J. Popielarski1, L. Popielarski1, E. Pozdeyev1, T. Russo1, K. Saito1, R. Webber1, J. Weisend1, M. Williams1, Y. Yamazaki1, A. Zeller1, Y. Zhang1, Q. Zhao1 1 Facility for Rare Isotope Beams, Michigan State University, East Lansing, MI 48824 USA 2 Thomas Jefferson National Laboratory, Newport News, VA 23606, USA 3 Physics Division, Argonne National Laboratory, Argonne, IL 60439, USA 4 Istituto Nazionale di Fisica Nucleare Laboratori Nazionali di Legnaro, Legnaro, Italy Abstract electrostatic deflectors for machine protection, a The Facility for Rare Isotope Beams, a new national Radiofrequency Quadrupole (RFQ) linac, linac segment 1 user facility funded by the U.S. Department of Energy (with Quarter-wave Resonators (QWR) of =0.041 and Office of Science to be constructed and operated by MSU, 0.085) accelerating the beam up to 20 MeV/u where the is currently being designed to provide intense beams of beam is stripped to higher charge states, linac segments 2 rare isotopes to better understand the nuclear physics, and 3 (with Half-wave Resonators (HWR) of =0.29 and nuclear astrophysics, fundamental interactions, and 0.53) accelerating the beam above 200 MeV/u, folding industrial and medical applications. The FRIB driver linac segments to confine the footprint and facilitate beam can accelerate all stable isotopes to energies beyond 200 collimation, and a beam delivery system to transport to MeV/u at beam powers up to 400 kW. Key technical the target a tightly focused beam. The reaccelerator (ReA) R&D programs include low- CW SRF cryomodules and consists of similar =0.041 and 0.085 accelerating highly efficient charge stripping using a liquid lithium structures [2]. film. Accelerator-physics challenges include acceleration Table 1: FRIB driver accelerator primary parameters. of multiple charge states of beams to meet beam-on-target requirements, efficient production and acceleration of Parameter Value Unit intense heavy-ion beams from low energies, Primary beam ion species H to 238 U accommodation of multiple charge stripping scenarios and ion species, designs for both baseline in-flight Beam kinetic energy on target > 200 MeV/u fragmentation and ISOL upgrade options, and design Maximum beam power on target 400 kW considerations of machine availability, tunability, reliability, maintainability, and upgradability. Macropulse duty factor 100 % 238 Beam current on target ( U) 0.7 emA INTRODUCTION Beam radius on target (90%) 0.5 mm The Facility for Rare Isotope Beams (FRIB), baselined as a 7-year, US$680 million construction project, is to be Driver linac beam-path length 517 m built at the Michigan State University under a corporate Average uncontrolled beam loss <1 W/m agreement with the US DOE [1]. FRIB driver accelerator is designed to accelerate all stable ions to energies above 200 MeV/u with beam power on the target up to 400 kW (Table 1). After production and fragment separation, the rare isotope beams can also be stopped, or stopped and then reaccelerated. The fast, stopped, and reaccelerated rare isotope beams serve a vast range of scientific users in the fields of nuclear physics and applications. As shown in Figure 1, the driver accelerator consists of Electron Cyclotron Resonance (ECR) ion sources, a low energy beam transport containing a pre-buncher and ___________________________________________ *Work supported by the U.S. Department of Energy Office of Science under Cooperative Agreement DE-SC0000661 #wei@frib.msu.edu Figure 1: Layout of the FRIB driver accelerator. 8 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 DESIGN PHILOSOPHY  Reliability: A Machine Protection System (MPS) Full-energy linac technology is chosen to deliver minimizes component damage and operational primary beam that can meet the FRIB requirements of interruption (e.g. magnet quench and cryogenic load rare-isotope productivity and separation accuracy. Up to increase) caused by both acute (fast) and chronic 400 kW of beams are focused to a diameter of 1 mm (slow) beam losses. Upon acute beam loss, the MPS (90%), energy spread of 1% (95% peak-to-peak), and response time is 35s (including diagnostics, signal bunch length of < 3 ns (95%) on the target. processing, and residual beam dumping). MPS In contrast to high-intensity spallation neutron sources responding to slow beam loss is complicated by low and neutrino sources that require pulsed beams, most of sensitivity of conventional ion chamber loss FRIB experiments prefer high-duty or continuous-wave detection to low-energy heavy ions and beam-loss (CW) beams. By choosing CW acceleration, a low beam signal background from adjacent linac segments with current (< 2 mA) can generate the required beam power which beam energies are significantly different. of 400 kW. Except for the ion source, effects of space Beam-halo- scraping rings in the warm interconnect charge are mostly negligible. sections and possible thermal sensors at cold regions Superconducting (SC) technology is the energy- are planned for more sensitive loss detection [4]. efficient choice for the CW linac. SC acceleration of  Tunability: The accelerator is designed to be easily heavy-ion beams is feasible from very low energy (500 tunable during both beam commissioning and keV/u) with practically sized cavity bores by housing operations [5]. In linac segment 1, where beam both the cavities and solenoids in a cryomodule. A two- transverse-phase advance is large, cold beam- cell scheme is chosen throughout the entire linac position monitors (BPM) are implemented in the providing both efficient acceleration and focusing. cryomodules. Efforts are made in establishing beam- Developments of digital low-level RF control and solid- tuning strategies based on virtual accelerators and state RF amplifier technologies have made individual on-line models under normal and fault conditions. cavity powering and control reliable and cost efficient.  Upgradeability: Space is reserved in linac segment 3 Furthermore, high availability, maintainability, to house another 12 cryomodules to readily increase reliability, tunability, and upgradability are especially the energy of 238U beam above 300 MeV/u. If required for the FRIB accelerator to operate as a national cavities with 35% higher accelerating gradient are scientific user facility. used in linac segments 2 and 3, the beam-on-target  Availability: The accelerator is designed with high energy can be raised above 400 MeV/u for 238U. The beam-on-target availability accommodating normal, linac tunnel allows future expansion so that a alternative, and fault scenarios. In the normal dedicated light-ion injector can be added supporting scenario, a liquid lithium stripper is used to raise the rare isotope production using the isotope separation average charge state of 238U beam to 78+ for efficient on-line (ISOL) method [1]. Using an RF deflector acceleration. Alternatively, helium gas confined by cavity and a Lambertson septum magnet, 3He+ beam plasma windows with differential pumping can be supplying protons to the ISOL target can share cycle used to strip the 238U beam to a lower average charge with the 238U beam feeding the fragmentation target; state of 71+. Fault scenarios include the situation thus simultaneous users are supported. Furthermore, when superconducting (SC) cavities underperform space is reserved to house instrumentations including by up to 20% of the design gradients. Furthermore, non-destructive diagnostics and sub-harmonic (e.g. key components (e.g., QWR cryomodules) and 20.125 MHz) buncher that are compatible with subsystems (e.g., machine protection) are future user demands of experiments. implemented with spares and design redundancies.  Maintainability: The average uncontrolled beam loss is limited to below 1 W/m level for all ion species from proton to uranium, to facilitate hands-on maintenance. For a proton beam at high energy, this level corresponds to an average activation of about 1 mSv/h measured at a distance of 30 cm from the beam chamber surface, 4 hours after operations shutdown [3]. For heavy ions like uranium at low energies, activation and radiation shielding is of less concern; the 1 W/m limit addresses concerns in damage on superconducting cavity surfaces and in cryogenic heat load. To facilitate maintenance of individual cryomodules, warm interconnect sections are used between cryomodules, and U-tubes with bayonet connections are used for cryogenic Figure 2: Layout of cryomodule inside FRIB driver distribution (Figure 2). accelerator tunnel showing U-tube cryogenic connections. Radioactive Ion Beams and Facilities 9

MOB01 Proceedings of HIAT 2012, Chicago, IL USA ACCELERATOR PHYSICS CHALLENGES TECHNOLOGY CHALLENGES The FRIB accelerator design combines the complexity Major R&D topics of FRIB accelerator systems include of heavy ion accelerators with the engineering challenges high-efficiency charge stripping and superconducting RF of high-power accelerators [3]. Due to the low charge-to- technologies for low-β acceleration. mass ratio, heavy ion acceleration is often not efficient. Uncontrolled beam loss, which usually is not an issue for Charge Stripper low-power heavy ion machines, is of primary concern for The FRIB baseline design of charge stripping [11] was the FRIB accelerator. Comparing with high-power proton based on the ANL work on a liquid-lithium high-power machines like the Spallation Neutron Source (SNS) linac thick target where it was demonstrated power deposition where apertures of the elliptical SC cavities [6] are large of 20 kW (from an electron beam) with the liquid lithium and beam amplitude reaches maxima in the warm operating well in an accelerator environment. During the locations of the focusing quadrupole magnets, the last couple of years the stability of a thin liquid film with apertures of the FRIB QWR and HWR accelerating the correct thickness for the FRIB stripper (~600 g/cm2) structures are small and beam amplitude reaches maxima was achieved (Figure 3). The next step is to show that the in the cold solenoid locations inside cryomodules. power deposited by the primary beam on the moving film Requirements on beam halo prevention, detection and (~ 1 kW) will not destroy or perturb it. With this purpose mitigation are stringent. we borrowed the LEDA ion source and LEBT from To maximize beam intensity on the target, beams of LANL and plan to modify the optics to obtain a 1 to 3 multiple charge states are accelerated simultaneously (2 mm diameter beam spot on the liquid-lithium film. As the charge states of 33+ and 34+ before stripping, and 5 protons will stop on the film the power will be charge states of 76+ to 80+ after stripping for 238U). comparable to the power deposited by the heavy ions Bends of second-order achromatic optics are used to fold during operation. The source reconfiguration is taking the beams, and cavity phases are adjusted so that beams place at MSU. Once the new optics is checked on a new are longitudinally overlapping at the charge stripper [7]. platform that is matched to the ANL lithium loop, the Conventional charge strippers like solid carbon foils are device will be moved to ANL for the integrated test. not sustainable at the power of a 238U beam at 17 MeV/u A second option for the stripper consisting of a helium during normal operations [8]. FRIB accelerator lattice gas cell enclosed by plasma windows was considered. needs to accommodate beam acceleration of different The RIKEN group has shown [12] that helium gas charge states resulting from various stripping methods produces a higher average charge state than a nitrogen gas including liquid lithium and helium gas (average charge stripper. The purpose of the plasma windows is to limit state from 63+ to 78+ for 238U). Buncher cavities of the helium gas leaking out of the gas cell [8]. fundamental (80.5 MHz and 322 MHz) and double (161 MHz) linac RF frequencies are strategically placed in the folding segments to preserve beam quality. Due to the short stopping distance in surrounding materials, uncontrolled beam loss of the low-energy heavy ions can cause damage to the surface of accelerating structures much more easily than a proton beam. On the other hand, due to the low level of radio- activation [9], losses of low-energy heavy-ion beams are difficult to detect [4]. Beam-loss detection and machine protection often rely on beam scraping. On the other hand, scraping of partially stripped ions may lead to higher ionization further complicating beam collimation and machine protection [10]. Due to requirements of frequent longitudinal and Figure 3: Liquid lithium film established at ANL. transverse focusing in the superconducting acceleration Superconducting RF structure, focusing solenoids are placed inside cryomodules adjacent to cavities. Alignment tolerance of FRIB driver linac is the first full-size SC linac using a these solenoids is ±1 mm under cryogenic conditions. large quantity (340) of low-β cavities. Cavity design is Horizontal and vertical steerers are needed to thread the optimized not only for optimum performance but also for beam and correct the beam orbit. low production cost [13]. This requirement guided the Stringent beam-on-target requirements demand tight choice of the cavity geometries, materials and mechanical optical control, error control, and advanced beam solutions, avoiding complicated shapes, minimizing the diagnostics. The primary beam of 400 kW needs to be amount of electron beam welds, eliminating bellows, focused into a diameter of 1 mm with below ±5 mrad optimizing construction and surface treatment procedures. transverse angular spread. The desired range of beam FRIB cavities work with superfluid helium at 2 K. The power variation on target is 8 orders of magnitude. Orbit increase in cavity Q more than compensates the loss of stability needs to be controlled at 0.1 mm level. efficiency of the 2 K cryogenic system. This innovative 10 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 choice in a low-β linac allows operation of cavities in Surface Treatments. Test results and production cost stable pressure conditions with high safety margin on the considerations led to the choice of buffered chemical maximum surface fields. polishing (BCP) for FRIB cavity’s surface treatment. The effort was concentrated in the development of a reliable Cavity Development. After a 10-year development, procedure [15] able to produce field-emission free, high the 2nd generation QWR prototypes are used in the ReA3 gradient cavities. The treatment includes the following linac (7 with β0=0.041 in operation and 8 with β0=0.085 BCP steps: 1) bulk etch (150µm removal), 2) differential under installation). This cavity type underwent etching in QWRs for final cavity tuning if required, 3) modifications including the displacement of the RF coupler from the bottom plate to the resonator side and an light etch (30 µm removal). Thermal treatment in high increased distance between the tuning plate and the inner vacuum at 600 °C is applied before step 3) for Hydrogen conductor tip, in order to remove a critical thermal removal to prevent Q disease. High pressure water rinsing problem in the design. The new tuning plate includes slots (HPR) is applied before cavity final installation. To assure and undulations to increase its maximum elastic surface cleanliness, dust particle count is performed on displacement and thus its tuning range, and a “puck” resonator surfaces during cavity assembly in the clean whose length can be adjusted for cavity tuning before room and the water purity is continuously monitored final welding. Concerning the β0=0.53 HWR prototype, during HPR. During BCP both the cavity and acid which is similar in design to the β0=0.29 cavity, 4 units of temperatures are stabilized to control the removal rate and the 2nd generation have been built by 2 different vendors to avoid excess of hydrogen absorption in Nb. The acid in addition to the one built in house. After positive test flow path in the β0=0.53 HWR was studied by means of results, the β0=0.085 QWR and the HWRs designs have simulations and experiments to obtain homogeneous been further refined in a 3rd generation, upgraded design removal over the entire inner resonator surface. Final with increased diameter that takes maximum advantage of thermal treatment at 120°C was also implemented, the space available in FRIB cryomodules (Figure 4 and showing significant reduction of the Q-slope at 4.2 K, but Table 2) [14]. negligible improvement at 2 K. Test Results in Prototypes. The 2nd generation prototypes of β0=0.085 QWR and 0.53 HWR have been tested in vertical dewars at 2 K exceeding FRIB design specifications (Figures 5 and 6) [14]. Figure 4: FRIB SC cavities. From left: β0=0.041 and 0.085 QWRs, β0=0.29 and 0.53 HWRs. Table 2: Upgraded FRIB cavity parameters Resonator QWR1 QWR2 HWR1 HWR2 Figure 5: ReA3, 0=0.085 QWR prototype Q0 vs. Eacc 0 0.041 0.085 0.29 0.53 at 2 K and 4.2 K, before and after 120°C baking. f (MHz) 80.5 80.5 322 322 V (MV) 0.81 1.8 2.1 3.7 a E (MV/m) 31 33 33 26 p B (mT) 55 70 60 63 p 9 Q0 (10 ) 1.2 1.8 5.5 7.6 R/Q (Ω) 402 452 224 230 G (Ω) 15 22 78 107 Aperture (mm) 34 34 40 40 Leff ≡  (mm) 160 320 270 503 Figure 6: Performance of the 0=0.53 HWR prototypes Number of cavities 12 94 76 148 Q0 vs. Eacc processed and tested at MSU and JLAB. Radioactive Ion Beams and Facilities 11

MOB01 Proceedings of HIAT 2012, Chicago, IL USA The residual resistance measured in the prototype Multipacting was found to impede the performance of the families was below 5 n up to about 100 mT in QWRs, cavities and the coupler. Future tests are planned starting and about 80 mT in HWRs. Considering that the FRIB August 2012 on magnetic field and shielding studies, specified limits are 11 n and 70 mT, a large safety mechanical-type and pneumatic-type tuner evaluations, margin exists for future upgrades (Figure 7). multipacting mitigation, and integrated tests with both β0=0.29 and β0=0.53 HWRs. ACCELERATOR DESIGN The FRIB facility is designed adjacent to the ReA facility [2] to benefit fully from the existing experimental infrastructure at the National Superconducting Cyclotron Laboratory (NSCL) (Figure 9). The driver linac is folded twice to minimize the footprint impact on campus, and the tunnel layout allows “open cutting” to save on civil construction costs. Design of the penetrations between the tunnel and the service buildings is compatible with the planned energy upgrade. Space is reserved to house kickers, septums, and target of a future ISOL facility and Figure 7: Residual surface resistance at 2 K of “stubs” are built to the tunnel walls to allow connection to 0=0.041 and 0.085 QWR and 0=0.53 HWR prototypes. a future light-ion injector for the planned upgrades. Design Upgrade. The 3rd-generation cavity design optimization resulted in significant improvement of peak fields Ep/Eacc, Bp/Eacc and shunt impedance Rsh, with consequent reduction of the overall linac cost and operational risk. Ep and Bp in operation could be moved below the safe values of 35 MV/m and 70 mT in all cavities. The design gradients of the β0=0.085 QWR and β0=0.29 HWR are raised by 10% without increasing the total cryogenic load, allowing a reduction of two cryomodules. The apertures of all QWRs were enlarged from 30 to 34 mm, and their bottom rings were modified for efficient tuning-plate cooling using a low-cost design [14]. The HWR designs were optimized to facilitate the mechanical construction and tuning procedure. In all cavities, the helium vessel is made of titanium to avoid Figure 9: Layout of the FRIB facility (colored) at MSU. brazed Nb-to-stainless-steel interface. Front End The FRIB Front End includes two ECR sources (ECRIS), two charge selection systems, a LEBT, a RFQ, and a MEBT (Figure 10). To enhance availability and maintainability, placed at the ground level in the support building about 10 m above the tunnel level are two ECR sources: a SC high-power source based on the SC ECRIS VENUS developed at LBNL [16] and, initially, a room- temperature ECRIS ARTEMIS available at NSCL. Figure 8: TDCM containing two 0=0.53 prototype HWR and a solenoid during installation in the test bunker. β0=0.53 Prototype Cryomodule (TDCM). The Technology Demonstration Cryomodule (TDCM) consists of two β0=0.53 HWRs operating at 2 K and a 9 T solenoid operating at 4.5 K arranged in a “top-down” configuration (Figure 8). The cryomodule was tested at the design cryogenic temperatures demonstrating excellent cryogenic and LLRF control stabilities. Figure 10: FRIB Front End with its major subsystems. 12 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 Figure 11: Layout of the FRIB driver accelerator at the tunnel level. The beams extracted from the sources are filtered in the Several potential “hot spots” are likely to compromise charge selection systems. The LEBT design is achromatic SRF performance: ion source, charge stripper, charge allowing transporting two charge states simultaneously to selector, and fragment target. Fast acting valves of ms double the accelerated beam intensity. To facilitate response time are designed at both ends of each linac transport of the two-charge state beams the LEBT uses segment to protect cryomodules against vacuum failures. electrostatic quadrupoles and two 90 dipole deflectors in Since energy gains in the cavity are different for beams the vertical transport line. To reduce losses in the of different charge states, a multi-reference-particle model superconducting linac, the beam is collimated by several is established to handle various charge states of respective apertures in the LEBT. Then, the CW beam is bunched by beam trajectory, beam energy, synchronous phase, focal a multi-harmonic buncher operating at 40.25, 80.5, and length and phase advance. Transverse coupling in a lattice 120.75 MHz before injection into the RFQ. A second RF containing solenoids presents another challenge to beam cavity operating at 40.25 MHz upstream of the RFQ is matching even though the beam is nominally round. required for two-charge-state beam injection acting as a Particle tracking simulation with 3D fields indicates that velocity equalizer and reducing the longitudinal beam even if a round beam is injected into the linac, it will emittance. A chopper is used to vary the duty cycle become tilted in the linac presumably due to the controlling the pulse length from several hundreds of ns quadrupole components of the QWR cavities. Transverse to CW and the pulse frequency from 0 to 30 kHz. Several matching of a coupled beam requires fitting of 10 mesh screens allow reduction of the beam intensity by variables comparing with 6 of a decoupled beam. several orders of magnitude while keeping the nominal bunch frequency. The two 90 electrostatic deflectors in Folding Segments and Beam Delivery System the vertical transport line are incorporated into the Two folding segments split the linac in three segments. machine protection system. The voltage on the deflector The stripper is located just upstream of the first 180˚ plates with shut-off time below 1 s are controlled by fast bend. Four 45˚ room-temperature dipoles and several switches that receive the inhibit signal from the MPS. combined function (quadrupole/sextupole) magnets The RFQ accelerates the beam from 12 to 500 keV/u. control the chromatic effects. Five charge states of 238U Downstream of the RFQ, the MEBT consists of two are transported with different rigidities (Q/Q ~ 3%). room-temperature QWR bunchers, four SC solenoids, an SRF rebunchers are located before and after the bend to energy analyzing dipole, and diagnostics. It matches the match the longitudinal phase space to the acceptance of beam to the SC linac and removes un-accelerated beam the second linac segment. This first bend effectively that is not longitudinally captured in the RFQ. facilitates charge selection and beam halo scraping. The second bend was introduced to reduce the tunnel Linac Segments footprint, as shown in Figure 11. Four superconducting Three sequential linac segments (LS1, LS2, and LS3), dipoles of 2 T maximum field adopt similar design to the each over 100-m long, accelerate the beam from 500 dipoles used at NSCL in the A1900 fragment separator. keV/u to above 200 MeV/u (Figure 11). There are 44 Afterwards, the beam is further accelerated in the third acceleration cryomodules and 5 rebuncher cryomodules segment and then bent 70˚ toward the production target in the linac containing a total of 330 QWR and HWR where the objective is to achieve 90% of the beam in a 1 cavities. 9-T SC solenoids provide beam transverse mm diameter beam spot with all five charge states focusing in all the acceleration cryomodules. In order not superimposed. The final component of the transport line to quench adjacent SC cavities, buckling coils are is the final focus triplet where special care is taken to equipped with each solenoid to limit the stray fields. reduce the chromatic aberrations. Meanwhile, residual magnetic fields are limited to no more than 15 mG to ensure SRF performance of the Reaccelerator cavities. Along with cryomodule magnetic shielding, an The re-accelerator facility (ReA) is a heavy ion linac automatic degaussing process is designed for the SC consisting of an Electron Beam Ion Trap (EBIT) charge solenoid. breeder, an off-line stable ion beam injector, a multi- Radioactive Ion Beams and Facilities 13

MOB01 Proceedings of HIAT 2012, Chicago, IL USA harmonic buncher, an RFQ, and 8 cryomodules (one assembly are planned to be performed in house, o=0.041 cryomodule, four o=0.085 cryomodules, and fabrication of a large quantity of repetitive components three bunchers) with a total of 41 SC cavities. ReA are planned through mass production and out-sourcing by reaccelerates the rare isotope ion beam after it was mass- industrial providers. Based on quoting and purchasing separated in the fragment separator and decelerated to experiences, a cost reduction of 2 to 3 times is expected thermal energies. The 1+ rare isotope beam is injected between prototypes and mass-produced items. into an EBIT charge breeder, stripped to a Q/A between The FRIB project plans to place approximately 450 0.2 and 0.5 and accelerated in the linac. ReA is designed procurements valued at more than $50k each. The sum of to provide beams with energies from 0.3 to 12 MeV/u for all technical equipment procurements amounts to $217M heavier ions and from 0.3 up to 20 MeV/u for light ions. excluding conventional facility construction. We have The first stage of ReA (ReA3) is partly under beam implemented a procurement strategy that strives to reduce commissioning. The first two o=0.041 cryomodules were vendor risks for the best value to the FRIB project: installed in 2010 and commissioned in 2011. The third  We work directly with the vendors understanding cryomodule (o=0.085) is planned to be installed by the their individual risk concerns and proposing end of 2012 completing the ReA3 project. In 2014 an mitigations: additional o=0.085 cryomodule will be added which will o Perform certain tasks in-house if the vendor enable acceleration of all ions above the Coulomb lacks capability; barriers. A limited user program utilizing rare isotopes o Adjust engineering designs to allow vendor to produced by the Coupled Cyclotron Facility (CCF) is implement familiar fabrication approaches; scheduled to start in 2013. o Identify key-personnel bottlenecks at the vendor, During the commissioning of the cryomodules, the and provide technical support where necessary; stabilities of SRF cavities and LLRF control were suc- o Work closely with the vendor in mass-production cessfully demonstrated to FRIB requirements [2]. Figure planning and subsequent quality monitoring; 12 shows the first beam energy spectrum observed in o Accept components on mechanical ReA3 during commissioning with a He+ beam. Each peak (dimensional) properties instead of functional corresponds to a beam energy shift after the RFQ when (e.g. electromagnetic, RF) performance; each cavity is sequentially turned on and phased for o FRIB to purchase high-cost materials that expose acceleration at nominal ReA3 gradients (Va=0.432MV). to market fluctuations (e.g., Nb). In addition, the first 1+ to n+ acceleration using the EBIT  We evaluate how the project fits into the supplier’s charge breeder was recently demostrated. For this test, a total capabilities and long-term business plans to single charged 39K beam was injected into the EBIT gauge supplier management’s commitment to solve source from the off-line test source, charge bred to 16+ production challenges and risks. and accelerated through the linac.  We develop long-term supplier relationships for mass production. Phasing of procurements from prototypes to production sensitizes vendors to be able to successfully produce the unique components. So far we have implemented this strategy to negotiated favorable prices for all the SRF material purchases and the production of 174 β=0.53 HRWs [17]. Cryogenics The cryogenics system is designed to support the operation of both SC cavities at a sub-atmospheric pressure (2 K) and SC magnets at a pressure slightly over atmospheric. The system must also provide a 4.5 K liquefaction load to support the magnet’s power leads and a non-isothermal refrigeration-shield load between 38 and 55 K. Table 3 summarizes the interface heat load requirements for each load temperature. Figure 12: Energy spectra measured with a silicon The distribution system consists of three separate linac- detector. By turning on one cavity at a time the nominal segment lines and a separator-area distribution line. It final energy of 5.486 MeV (1.38MeV/u) was reached. uses cryogenic disconnects consisting of vacuum insulated “U-tubes” and bayonets integrated with non- SUBSYSTEM DESIGN & ACQUISITION cryogenic isolation valves that are similar to JLab and Design & Acquisition Strategy SNS designs. Each segment, as well as individual cryomodules, may be cooled down and warmed up After the project baseline, the FRIB project entered into independently. its final design phase when detail engineering designs are The options to support various cryomodule loads are performed at MSU. While critical processing and described and evaluated in [18]. The number and location 14 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 of valves and interface locations between distribution and coldmass assembly are planned to be performed in- system and cryomodules are considered to accomplish house. The cryomodule thermal shield, magnetic shield, cool down, warm up, and isolation, as well as for efficient cryogenic plumbing, and vacuum vessel will be procured and steady operations. from industry as complete “plug-in” units ready for Since the dominant refrigeration load is at 2 K, the assembly in the cryomodules. refrigeration process is based on the process options study presented in [19] and incorporates the cumulative experience from both JLab and SNS cryogenic systems. Since the load estimates are approximate, margins are included in these load estimates for the refrigerator. The recent experience gained from the JLab 12 GeV cryogenic system design is utilized for both the refrigerator cold box and the compression system designs. The Floating Pressure Process – Ganni Cycle [20] is to be implemented to provide efficient adaptation to the actual loads. Table 3: Interface Load Requirements. In addition, the total magnet lead flow is 3.2 g/s, representing 3% of the total load exergy. Source Heat load [W] 2K 4.5 K 38 – 55 K Cryomodule 2490 1470 7690 Magnets 0 670 1000 Figure 13: FRIB “bottom-up” cryomodule designs. The Cryodistribution 0 950 5000 top cryomodule incorporates 8 β=0.085 QWRs, 3 Beam loss 0 25 0 solenoids, and 3 cold beam position monitors; the bottom cryomodule incorporates 8 β=0.53 HWRs and a solenoid. Total load 2490 3115 13690 Load exergy fraction 54% 30% 13% RF The RF System monitors and controls the amplitude Cryomodule and phase of the voltage in the cavities. The RF System consists of the reference clock generation and distribution After the conceptual design, the FRIB cryomodule line, LLRF controllers, RF amplifiers, RF transmission design has evolved significantly from the “top-down” lines including couplers and pickup cables, cavity spark ReA3 style to a “bottom-up” design. Key features detectors, and cavity tuner controls (Figure 14). including rail system, support system, heat and magnetic shields are simplified along with improvements in assembly and alignment. Figure 13 shows the current design incorporating a torque-resistant structural frame made of stainless steel. We incorporate machined fiberglass compression posts supporting the coldmass in the cryomodule vacuum vessel. Three posts on linear roller bearings oriented towards the center of thermal contraction serve as 6-degree-of-freedom kinematic supports. This design controls the alignment of the coldmass while allowing thermal contraction. Cavity and solenoid attachment points to the rails are all machined after welding to ensure assembly consistency [21]. The FRIB cryomodules contain separate cryogenic circuits at 4 K for the solenoids and at 2 K for the cavities with the 2 K heat exchanger residing inside. We are Figure 14: RF system interface block diagram. currently evaluating cavity operation at 2.1 K. The slight reduction in cavity Q would be offset by substantial The RF Reference clock is generated by a 10.0625 savings in cryoplant costs. However, care must be taken MHz oven-controlled crystal oscillator and distributed along the linac via phase-stabilized RF cable with for quiescent cavity operation close to the helium -point. temperature variation below ±5ppm/˚C. The reference The project schedule calls for a production rate of 2 cryomodules per month. Cavity processing, vertical tests line has directional couplers near each cryomodule. The Radioactive Ion Beams and Facilities 15

MOB01 Proceedings of HIAT 2012, Chicago, IL USA reference signal cable from the directional coupler to the transmission efficiency above 80%. It operates at 80.5 LLRF module in the service building is routed as close as MHz and requires about 100 kW of CW RF power. A 4- possible to the forward, reflected, and cavity feedback vane resonator is utilized with a liner ramped accelerating signals to minimize the effects of temperature drift. voltage profile (Figure 15). The LLRF controller directly samples the cavity feedback signals and the reference signal. The sample clock is locked to and derived from the reference clock. The sample clock frequency is chosen to minimize the aliasing effects of harmonics that may be present in the incoming signals. The raw samples are filtered and converted to phase and amplitude in the field- programmable gate array (FPGA) and control logic directly produces the output frequencies using a high- speed digital-to-analog converter (DAC). The tuner control signals are also generated in the FPGA using information from the RF feedback signals. A new control method - active disturbance rejection control (ADRC), is applied to solve the problem of microphonics detuning in the SC cavities. ADRC Figure 15: RFQ assembly view of segment 1 with provides 200% improvement over the proportional- endplate removed. integral-derivative (PID) method in simulations and 400% improvement in hardware tests. A digital self-excited Mechanical construction of the RFQ will be performed loop mode was also implemented in the LLRF controllers. as an integral brazed structure with dual-circuit cooling With the exception of the RFQ amplifier (200 kW water resonance control. Overall mechanical fabrication tetrode), the FRIB RF amplifiers use solid-state is performed in 5 longitudinal sections to minimize technology. Multiple amplifiers can be combined to machining and handling weight. The quadrupole and generate the required power (up to 8 kW for =0.53 dipole mode frequencies are fine-tuned during HWRs). The isolation provided by the circulators allows construction using fixed mechanical slug tuners in all 4 the use of N-way power combiners which saves space and quadrants distributed along the length of the machine. cost. The high-power RF cables, connectors, circulators, Dipole mode suppression rods attached to the structure power combiners and transmission lines for the SRF endplates are utilized to separate and move dipole mode amplifiers are rated to handle the standing waves frequencies away from the accelerating mode frequency, produced by the reflected power due to over-coupling. thus stabilizing the desired quadrupole accelerating mode Ion Source at 80.5 MHz. The field ramp in the accelerating mode is accomplished through proper sizing of the magnetic The ARTEMIS ECR source built at MSU based on the return vane undercuts at the ends of the structure and AECR-U design at LBNL operates at 14.5 GHz with power is fed using a magnetic-field loop-coupler drive. room-temperature coils. Thermal management of CW RFQ’s is challenging but The VENUS-style ECR source will be a SC high- given the modest drive power of below 100 kW, no performance ECR source operating at 28 GHz. In 2007, significant cooling issues exist. All components are made VENUS demonstrated the FRIB current [16] for a 238U of high-conductivity copper and actively water cooled: beam when combining the two charges states of 33+ and resonator vanes, endplates, slug tuners, vacuum pumping 34+. Recent beam measurements demonstrated that better port grills and the 15.6-cm diameter coaxial drive coupler performances from VENUS can be obtained by coupling power feed. Additional care is taken in the vane-undercut additional microwave power (up to 8 kW) to the ion regions where the local RF heating is increased due to source. A new intensity record of 450 eA of 238U33+ was magnetic-field compression in these zones. obtained with VENUS at LBNL. The measured emittance for this charge state showed that 95% of the beam was Magnet and Power Supply within the FRIB acceptance. FRIB modifications to the Most of the magnets in the driver linac are resistive due original design of the VENUS cryostat include the to the relatively low beam rigidity. Superconducting cooling capacity at 4.2K been extended from 6 W to 12 W dipoles are used when fields of 2 T are required (folding to include the dynamic heat load generated by the plasma segment 2). The fragment separator uses SC technology Bremsstrahlung electrons. to allow the high pole-tip fields and large apertures Both sources are installed on a high-voltage platform to required to maximize the acceptance. Re-configuration of provide an initial beam acceleration of 12 keV/u at FRIB. the existing A1900 fragment separator allows connection RFQ to the existing NSCL experimental equipment. Power supplies are largely in three categories, room- The FRIB RFQ accelerates the multi-charge state beam temperature magnet power supplies, SC magnet power from 12 to 500 keV/u over a 5-m distance with estimated 16 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 supplies, and high-voltage power supplies. Programmable Conceptual designs for fixed- and adjustable-aperture power supplies provide regulated DC current to the halo rings are shown in Figure 16. magnets, or regulated high voltage to electrostatic elements and ion source components. Vacuum and Alignment Beam-boxes as shown in Figure 16 of two standardized Diagnostics lengths and identical diagnostics ports will be installed The driver linac is planned to be commissioned with across the entire driver accelerator. Over 100 ion pumps single-shot or low repetition rate, 50 µA and 50 µs beam are used to generate vacuum ranging from 10-6 Torr pulses. This sets the fundamental operating specification (charge-stripping section) to 5x10-9 Torr (warm region for most diagnostics systems. In full operation, the between cryomodules). Vacuum requirements are diagnostics must accommodate low-to-moderate energy, primarily determined by the need to minimize stripping of low instantaneous current, CW, high-power beams of the highly-charged heavy ions. Effects of electron cloud differing ion species with high charge states. are expected to be negligible. The Molflow+ codes [22] Use of intercepting diagnostic devices is acceptable are used to guide the pump sizing and placements. only under tightly controlled, low-beam-power conditions Planning for the alignment of the facility is underway. and only at locations not adjacent to SC cavities. Also, In the near-term we focus on designing the alignment FRIB low-energy, heavy-ion beams do not avail network including characterization of the existing themselves to many optical diagnostics techniques used network in the NSCL facility matching the FRIB global- for highly-relativistic beams. Beam position and beam coordinate system. The alignment network encompasses current monitoring system design is challenged by the the front end beam lines at ground level, a 10-m vertical continuous, low-current (500 µA maximum) nature of the drop to the over 3,066 m2 linac tunnel, heavily shielded FRIB beam. The present specification for the operational target system chambers, and a vertical rise back to and “CW” beam includes a 50 µs beam gap at 100 Hz including the entire existing NSCL facility. primarily for diagnostics purposes. Weight and structural analysis and differential- The FRIB beams present particular challenges for beam settlement monitoring is included as part of conventional- loss monitoring, which is traditionally expected to facility design. provide important signals for machine protection and for beam tuning and optimization. Full loss of the CW heavy Controls ion beam, even at relatively low energies, can cause As the interface for operators and physicists to component damage in tens of microseconds. At low commission and operate the FRIB accelerator, the control energies, e.g. all of LS1, the radiation signature outside system provides supervisory and model-driven control the beam line elements can be vanishingly small. (Table 4). Distributed, chronic, fractionally small losses of the CW beam can adversely load the cryogenic system. Radiation Table 4: FRIB distributed control system. cross-talk between adjacent segments of the folded FRIB Physical Distribution ~200 m * 200 m linac configuration is large [4]. Finally, there is expected IOCs ~150 Computers/EPICS to be a competing radiation field due to X-rays from SRF Input/Output Controllers (IOC) cavities, which is yet to be quantified. PLCs ~100 Programmable Logic Controllers (PLC) Network ~ 3000 GbE ports Global Timing >500 timing drop points Machine Protection ~2000 MPS fast protection inputs Conventional Facility ~760 racks and more than 100 with controls devices The low-level control system consists of programmable-logic-controllers (PLC), input-output- controllers (IOC) and signal-conditioning modules to provide process control and remote operation to the field devices including vacuum devices, power supplies, RF controllers and cryogenic sensors/controllers. Figure 16: Standardized accelerator vacuum chambers The global system provides site-wide timing (event and with diagnostics ports. Also shown are fixed (left) and data), network connectivity and integration of machine variable (right) diameter halo scrapers for interceptive protection to allow all subsystems to work together with beam-loss monitoring. synchronized real-time information, Ethernet-based communication and consistent fault handling. The Aperture-limiting “halo rings”, instrumented to sense machine mode, beam mode and particle type are first set the deposited charge, are being considered for monitoring in global system and then distributed throughout the chronic, small beam losses [4]. The halo rings would be facility to ensure that every related subsystem has located in the warm, inter-cryomodule regions. consistent information. Radioactive Ion Beams and Facilities 17

MOB01 Proceedings of HIAT 2012, Chicago, IL USA The high-level applications provide both web- and LCW system supplies the Front End systems. These are console-based operation toolkits, relational-database nominally 3-M systems at 32˚C supply temperature services and model-based physics applications to support incorporating oxygen reducing system and ultra-violet installation, commissioning and operations. The service- (UV) control of biological growth. Service building rack oriented architecture is selected, the key technologies cooling is accomplished using filtered, treated water since such as Eclipse/CSS, J2EE, XAL and MySQL are decided LCW is not required. and a set of tools are developed. Instrument-grade compressed air is routed throughout the facility for the operation of control valves and Utilities solenoids, as well as for blowing out water lines and A robust infrastructure is designed to support FRIB magnets during maintenance. accelerator operations including facility power, cooling Heating, ventilation, and air-condition (HVAC) system water, air, cable trays and conduits, facility layout and loads and type are specified for all areas of the facility. grounding. Fan-coil units which are supplied by hot and cold water The FRIB facility will be served by two 25 MW, 13.8 systems are located throughout the accelerator tunnel. kV primary feeders from the MSU T.B. Simon Power These units provide the primary air temperature Plant using an automatic transfer arrangement to transfer management during operation. A negative (5 mbar sub- load in the event of a circuit failure. The feeders are atmospheric) tunnel pressure during operation is provided installed in an underground duct bank providing via a 142 m3/min exhaust system venting through an protection from foreseeable hazardous weather. An exhaust gas management system with make-up air being additional 4 MW, 13.8 kV primary feeder is also routed to pre-conditioned through a conventional HVAC unit. the site to serve as investment protection in the unlikely Service building rack area cooling is via a raised floor scenario that both primary feeders are lost. This 4-MW “server room” style cooling system. Air will be supplied feeder requires manual operation to energize and serve at 13˚C through the floor and ducted to racks as needed. loads associated with cryogenics operations. These three Ground vibration may induce movement in the SC circuits serve a line-up of medium-voltage 15 kV switches cavities, disrupting the beam and demanding increased in a loop configuration with multiple tie-switches offering RF power. The warm-gas compressors of the helium flexibility in transporting power to various substations refrigeration system are a major concern, especially since throughout the facility. Critical loads that cannot tolerate the system is planned to be housed adjacent to the service a power failure such as control equipment for cryogenics, building above the tunnel to save conventional- communications, oxygen-deficiency-hazard (ODH) construction costs. Studies concluded that the compressor monitoring, personnel and safety protection systems are vibration can be adequately mitigated by mechanically served by three 200 kW (N+1) uninterruptible power isolating compressors using commercially-available supplies (UPS). The UPS’s will maintain power to the foundation pads [24]. The effects from other equipment critical loads while two 800-kW diesel generators like cooling-water pumps and chillers also are assessed operating in parallel come on-line to provide power to the and comparable vibration-reduction measures critical, emergency, legally required standby, and optional implemented. standby loads and the fire pump. A facility grounding plan was developed that minimizes conducted and FUTURE PERSPECTIVES radiated electromagnetic interferences. Ground mesh and The FRIB accelerator design is advanced towards mechanical bonding are strategically located to optimize beginning construction in 2014. Early procurements effectiveness [23]. before 2014 is strategically planned to establish the front FRIB driver accelerator’s electrical components will be end test stand to demonstrate critical components of the housed in 768 racks in the service building. The racks are ECR ion source and the fully powered RFQ, to contract organized in 120 groups that follow the contour of the on long lead-time cryogenic refrigeration subcomponents, accelerator. Conduit banks are located near each rack to acquire cryogenic distribution equipment that needs to group to keep cable lengths as short as possible. There be installed before accelerator tunnel completion, and to are 1200 conduits in these banks that carry electrical acquire SRF components to be assembled in the pre- signals from RF, power supplies, diagnostics, and controls production cryomodules. Upon fabrication, installation, in the accelerator tunnel. The conduits are 6-m long, and and integrated tests, early beam commissioning is nominal cable lengths are about 30-m long. Thermal expected to be staged from 2017 to 2019. The facility is modelling is being performed for cable heating in the scheduled to meet key performance parameters conduits. The conduits are organized for each rack supporting routine user operations before 2021. Full grouping by RF, DC power supply leads, high voltage, design capability is expected to be reached within 4 years controls, personnel-protection system (PPS), and AC after the beginning of routine operations. Science driven power distribution for cryogenic heaters. upgrade options may be pursued at any stage of the Three primary water cooling loops are specified for the project. accelerator systems’ technical equipment. An activated low-conductivity-water (LCW) system supplies the equipment in the accelerator tunnel while a non-activated 18 Radioactive Ion Beams and Facilities

Proceedings of HIAT 2012, Chicago, IL USA MOB01 ACKNOWLEDGMENTS [7] X. Wu et al, “The Overview of the Accelerator FRIB accelerator systems design has been assisted System for the Facility for Rare Isotope Beams at under work-for-others agreements by many national Michigan State University, LINAC’10, Tsukuba, p. laboratories including ANL, BNL, FNAL, JLAB, LANL, 163; Q. Zhao et al, “Beam Dynamics in the FRIB LBNL, ORNL, and SLAC, and in collaboration with Linac”, HB’12, Beijing (2012, to be published). many institutes including BINP, KEK, IMP, INFN, INR, [8] F. Marti et al., “Development of Stripper Options for RIKEN, TRIUMF, and Tsinghua University. We thank FRIB”, LINAC’10, Tsukuba, p. 662. the Accelerator Systems Advisory Committee chaired by [9] R. Ronningen et al, Nucl. Tech., 168 (2009) 670. S. Ozaki for their valuable advice and guidance, B. [10] T. Roser, private communications. Laxdal, P. Kneisel, and A.C. Crawford for their [11] Y. Momozaki et al., JINST 4 (2009) P04005. participation in the FRIB weekly teleconference, [12] H. Okuno et al., Phys. Rev. ST Accel. Beams, 14, colleagues who participated in FRIB accelerator peer (2011) 033503. reviews including A. Aleksandrov, G. Ambrosio, W. [13] C. Compton et al., “Superconducting Resonators Barletta, G. Bauer, G. Biallas, J. Bisognano, S. Bousson, Production for Ion Linacs at Michigan State M. Champion, M. Crofford, R. Cutler, B. Dalesio, G. University”, SRF’11, Chicago (2012, in press). Decker, J. Delayen, H. Edwards, J. Error, J. Fuerst, K. [14] A. Facco et al, “Superconducting Resonators Kurukawa, J. Galambos, J. Galayda, J. Gilpatrick, S. Development for the FRIB and ReA Linacs at MSU: Gourlay, S. Henderson, L. Hoff, G. Hoffstaetter, J. Recent Achievements and Future Goals”, IPAC’12; Hogan, N. Holtkamp, H. Horiike, C. Hovater, H. Imao, R. J. Popielarski et al., “Dewar Testing of Coaxial Janssens, R. Keller, J. Kelley, P. Kelley, M. Kelly, J. Resonators at Michigan State University”, IPAC’12, Kerby, A. Klebaner, J. Knobloch, J. Mammosser, T. New Orleans (2012, in press). Mann, N. Mokhov, G. Murdoch, H. Okuno, P. [15] L. Popielarski et al., “Cleanroom Techniques to Ostroumov, R. Pardo, S. Peggs, T. Peterson, C. Piller, J. Improve Surface Cleanliness and Repeatability for Power, T. Powers, J. Preble, D. Raparia, T. Roser, R. SRF Coldmass Production”, IPAC’12, New Orleans Ruland, W.J. Schneider, D. Schrage, J. Sondericker, W. (2012, in press). Soyars, C. Spencer, R. Stanek, M. Stettler, J. Stovall, V. [16] D. Leitner et al, Rev. Sci. Instru. 79 (2008) 02C710. Verzilov, P. Wanderer, M. Wiseman, L. Young, and A. [17] M. Leitner, “Design Status of the SRF Linac Systems Zaltsman, and colleagues who advised and collaborated for the Facility for Rare Isotope Beams”, SRF’11, with the FRIB team including A. Burrill, K. Davis, W. Chicago (2011, in press). Hartung, A. Hutton, R. Kersevan, S.H. Kim, K. Macha, [18] V. Ganni, “Helium Refrigeration Considerations for G. Maler, E.A. McEwen, W. Meng, T. Reilly, J. Cryomodule Design,” Proc. ADS Conf., Mumbai Sandberg, J. Tuozzolo, and J. Vincent. At Michigan State (2011, in press). University, the FRIB accelerator design is executed by a [19] P. Knudsen, V. Ganni, “Process Options for Nominal dedicated team of the FRIB Accelerator Systems Division 2-K Helium Refrigeration System Designs”, Adv. with close collaboration with the Experimental Systems Cryo. Eng. 57, AIP, New York (2012, in press). Division headed by G. Bollen, the Conventional Facility [20] V. Ganni, P. Knudsen, “Optimal Design and Division healed by B. Bull, the Chief Engineer’s team Operation of Helium Refrigeration Systems Using headed by D. Stout, and supported by the project controls, the Ganni Cycle,” Adv. Cryo. Eng. 55, AIP, New procurements, ES&H of the FRIB Project, by the NSCL, York (2010) p. 1057. and by the MSU. [21] M. Johnson et al, “Design of the FRIB Cryomodule”, IPAC’12, New Orleans (2012, in press). REFERENCES [22] Codes Molflow+, R. Kersevan, CERN. [1] Web site http://www.frib.msu.edu/ [23] M. Thuot, “Using the FRIB Facility Grounding Mesh [2] D. Leitner et al, “Status of the ReAccelerator Facility Effectively”, FRIB Internal Document T10503-VS- RεA for Rare Isotopes beam research”, SRF’2011, 000001 (2012); D. Stout et al, “201105 Grounding Chicago (2011, in press). Workshop Report”, FRIB Internal Document [3] J. Wei, Rev. Mod. Phys., 75, (2003) 1383. T20100-RA-000043 (2011). [4] Z. Liu, Y. Zhang et al, “Ion Chambers and Halo [24] H. Amick, “Review of Ground-Borne Vibration Rings for Loss Detection at FRIB”, IPAC’12, New Issues”, FRIB Internal Document T20201-TD- Orleans (2012, in press). 000082 (2011); J.A. Moore, E. E. Ungar, FRIB [5] Y. Zhang, “Beam Tuning Strategy of the FRIB Linac Internal Document T20201-TD-000083 (2011); D. Driver”; P. Chu et al, “Online Physics Model Stout et al, “FRIB Ground-Born Vibration Studies”, Platform”, IPAC’12, New Orleans (2012, in press). FRIB Internal Document T31200-TD-000084 (2011). [6] N. Holtkamp, “The SNS Linac and Storage Ring: Challenges and Progress Towards Meeting Them”, EPAC’02, Paris, p. 164. Radioactive Ion Beams and Facilities 19


Document Created: 2016-05-11 19:15:49
Document Modified: 2016-05-11 19:15:49
© 2025 FCC.report
This site is not affiliated with or endorsed by the FCC